A rate-distortion framework is used to define a very lowareas of mouth and eyes of videophone sequences. Consequently, bit-rate coding scheme based on wireframe model adaptation and more bit rate is allocated to coding of motion failure areas. The optimized selection of motion estimators. This technique achieves source of the problem is that motion measurements are obtained maximum reconstructed image quality under the constraint of a target on a sparse grid of points, which may not be representative of bit rate for the coding of the vector field and the wireframe representathe motion on arbitrary locations. Furthermore, measurements are tion information. First, a complete scheme is proposed for hybrid twoobtained independently from each other, and may result in an dimensional (2D) and 3D motion estimation and compensation. The overdetermined motion model in the case of very small or simple wireframe adaptation and updating is optimized for hybrid motion motion. estimation in the rate distortion sense. A more sophisticated tech-More sophisticated techniques are required to maintain the nique, adapted to the requirements of a very low-bit-rate coder is also proposed which considers also the transmission of the prediction distortion below the level of annoyance. These methods are comerror corresponding to the particular choice of the motion estimator monly based on extraction and coding of moving objects in the for each object in the scene. Experimental results illustrating the perscene [4-8]. The performance of such techniques is based on formance of the proposed techniques in very low-bit-rate image sethe success of segmentation and modeling of moving objects, quence coding application areas are presented and evaluated. α§ 1998
and therefore they may fail with complex scenes. Affine two-